Rotation direction determination system and method for a rotating body

Abstract

This invention provides a rotation direction determination system and method for quickly and accurately detecting and determining the rotation direction of a rotating object in an image while suppressing the load on the PC. [Solution] The rotation direction determination system 100 for a rotating body includes: an imaging means for photographing the rotating drum of a mixer truck, which is the rotating body to be determined; an image acquisition means for acquiring multiple images at predetermined frame intervals from the captured video information; a detection region setting means for setting parts on the surface of the rotating body that have characteristics in the multiple images as detection regions to be used for detection; a continuous image creation means for extracting the detection regions set by the detection region setting means from the multiple images and combining them in chronological order to create a continuous image; a feature quantity calculation means for calculating a Hog feature quantity, which is a feature quantity that is a histogram of the gradient direction of brightness, from the continuous image created by the continuous image creation means; and a rotation direction determination means for determining the rotation direction of the rotating body based on the Hog feature quantity calculated by the feature quantity calculation means.

Description

Technical Field

[0001] The present invention relates to a rotating body rotation direction determination system and a determination method that can quickly and accurately detect and determine the rotation direction of a rotating body of a rotating machine, particularly the drum of a mixer truck that transports fresh concrete, by using video information captured by a photographing means such as a surveillance camera.

Background Art

[0002] Conventionally, mixer trucks that transport fresh concrete manufactured at a fresh concrete manufacturing plant to a construction site have been used. When loading fresh concrete into the drum, it is necessary to rotate the drum and stir it. Also, during transportation, it is transported while rotating and stirring the rotating drum. Generally, the drum of a mixer truck rotates clockwise during transportation to stir the fresh concrete and is discharged by counterclockwise rotation. If the drum of the mixer truck is operated in a reverse rotation state during the loading operation, the fresh concrete may be accidentally discharged.

[0003] Conventionally, a drum rotation direction confirmation device that can easily confirm the rotation direction of a rotating drum from the driver's seat has been proposed (Patent Document 1). The drum rotation direction confirmation device described in Patent Document 1 is provided with a lever angle detection means for detecting the operation angle of a lever in the driver's seat that controls the rotation direction of the rotating drum, and a drum rotation direction index means for indicating the rotation direction of the rotating drum in the lever operation state according to the operation state of the lever obtained by the lever angle detection means.

[0004] Also, in order to detect the rotation direction of the rotating drum, a method of providing a detection marker on the surface of the rotating drum in advance and detecting the rotation direction using optical flow can be considered. In this case, as prerequisite conditions, (1) the brightness (luminance distribution) before and after movement is unchanged. (2) The image is smooth (differentiable temporally and spatially). (3) The amount of movement is extremely small.

Prior Art Documents

Patent Documents

[0005] [Patent Document 1] Utility Model Registration No. 3141681 Gazette [Overview of the project] [Problems that the invention aims to solve]

[0006] However, the drum rotation direction confirmation device described in Patent Document 1 allows confirmation from the driver's seat, but if the driver is not in the driver's seat during loading, there is a problem that the concrete may be mistakenly discharged if the drum of the mixer truck is rotating in the reverse direction during loading.

[0007] Furthermore, since the method of detecting the direction of rotation using optical flow requires the above-mentioned prerequisites to be met, it is difficult to detect the direction of rotation using video information captured by imaging means such as surveillance cameras. In addition, when using optical flow, the image processing is computationally intensive and requires a dedicated image processing device, resulting in a system configuration that is specialized and relatively expensive.

[0008] Therefore, the present invention solves the above-mentioned problems of the prior art, and its objective is to provide a rotation direction determination system and method for a rotating body that can quickly and accurately detect and determine the rotation direction of a rotating body shown in an image while suppressing the load on the computer. [Means for solving the problem]

[0009] According to the present invention, a rotation direction determination system for a rotating body comprises: an imaging means for imaging a rotating body to be determined; an image acquisition means for acquiring a plurality of images at predetermined frame intervals from video information captured by the imaging means; a detection region setting means for setting portions having characteristics on the surface of the rotating body as detection regions to be used for detection in the plurality of images acquired by the image acquisition means; a continuous image creation means for extracting the detection regions set by the detection region setting means from the plurality of images and combining them in chronological order to create a continuous image; a feature quantity calculation means for calculating a Hog feature quantity, which is a feature quantity that is a histogram of the gradient direction of brightness, for the continuous image created by the continuous image creation means; and a rotation direction determination means for determining the rotation direction of the rotating body based on the Hog feature quantity calculated by the feature quantity calculation means (hereinafter referred to as "the first rotation direction determination system for a rotating body of the present invention").

[0010] According to the first rotation direction determination system of the present invention, the rotation direction of a rotating body to be determined is determined by photographing the rotating body to be determined by the photographing means, multiple images are obtained from the captured video information by the image acquisition means at predetermined frame intervals, in the multiple acquired images, the portion having a characteristic on the surface of the rotating body is set as a detection region to be used for detection by the detection region setting means, the detection regions set by the detection region setting means are extracted from the multiple images, and a continuous image creation means is used to combine them in chronological order to create a continuous image, a feature amount calculation means calculates a Hog feature amount, which is a feature amount that is a histogram of the gradient direction of brightness, from the created continuous image, and the rotation direction determination means determines the rotation direction of the rotating body based on the calculated Hog feature amount, thereby enabling rapid and accurate detection and determination of the rotation direction of a rotating body captured in an image while suppressing the load on the computer, and can be implemented at low cost.

[0011] In the first rotation direction determination system of the present invention, the rotating body is a rotating drum of a mixer truck that transports ready-mixed concrete, the detection area setting means is configured to draw at least one temporary straight line parallel to the rotation direction at a position that includes a characteristic portion on the surface of the rotating body as the detection area, and set it as the detection area, the continuous image creation means is configured to extract images on the at least one temporary straight line and combine them in chronological order to create at least one continuous image, and the feature quantity calculation means is preferably configured to calculate the Hog feature quantity for the at least one continuous image (hereinafter referred to as the "second rotation direction determination system of the present invention").

[0012] The second rotating body rotation direction determination system of the present invention can determine the rotation direction of the rotating drum of a mixer truck. This prevents loading operations from being performed while the drum is rotating in the reverse direction.

[0013] In the second rotation direction determination system of the present invention, the detection area setting means is configured to draw a plurality of temporary straight lines parallel to the rotation direction at positions including characteristic parts on the surface of the rotating body as detection areas, and set them as detection areas; the continuous image creation means is configured to extract images on the plurality of temporary straight lines and combine them in chronological order to create a plurality of continuous images; the feature amount calculation means is configured to calculate the Hog feature amount for each of the plurality of continuous images; and the rotation direction determination means is preferably configured to determine the rotation direction of the rotating body based on the Hog feature amounts of each of the plurality of continuous images, and to determine the rotation direction of the rotating body by majority vote based on the respective determination results (hereinafter referred to as the "third rotation direction determination system of the present invention").

[0014] According to the third rotation direction determination system of the present invention, multiple temporary straight lines are drawn and set as detection areas, images on each of the multiple temporary straight lines are extracted and combined in chronological order to create multiple consecutive images, and the rotation direction of the rotating body is determined based on the Hog features of each of the multiple consecutive images, thereby enabling more accurate detection and determination of the rotation direction of the rotating body.

[0015] According to the present invention, a method for determining the rotation direction of a rotating body comprises: a shooting step of photographing the rotating body to be determined; an image acquisition step of acquiring a plurality of images from the captured video information at predetermined frame intervals; a detection region setting step of setting a portion having a characteristic on the surface of the rotating body in the acquired plurality of images as a detection region to be used for detection; a continuous image creation step of extracting the set detection regions from the plurality of images and combining them in chronological order to create a continuous image; a feature calculation step of calculating a Hog feature quantity, which is a feature quantity that is a histogram of the gradient direction of brightness, from the created continuous image; and a rotation direction determination step of determining the rotation direction of the rotating body based on the calculated Hog feature quantity (hereinafter referred to as "the fourth method for determining the rotation direction of a rotating body of the present invention").

[0016] The fourth method for determining the rotation direction of a rotating body according to the present invention allows for the rapid and accurate detection and determination of the rotation direction of a rotating body captured in an image while suppressing the load on the computer. Furthermore, it can be implemented at low cost.

[0017] In the fourth method for determining the rotation direction of a rotating body of the present invention, the rotating body is a rotating drum of a mixer truck that transports ready-mixed concrete, and in the detection area setting step, at least one temporary straight line is drawn in a direction parallel to the rotation direction at a position that includes a characteristic portion on the surface of the rotating body as the detection area, and this is set as the detection area, in the continuous image creation step, images on the at least one temporary straight line are extracted and combined in chronological order to create at least one continuous image, and in the feature amount calculation step, the Hog feature amount is calculated for the at least one continuous image (hereinafter referred to as "the fifth method for determining the rotation direction of a rotating body of the present invention").

[0018] According to the fifth method for determining the rotation direction of a rotating body of the present invention, the rotation direction of the rotating drum of a mixer truck that transports ready-mix concrete can be determined. This prevents loading operations from being performed while the drum is rotating in the wrong direction.

[0019] In the fifth method for determining the rotation direction of a rotating body according to the present invention, it is also preferable that in the detection area setting step, a plurality of temporary straight lines are drawn in a direction parallel to the rotation direction at positions including a characteristic portion on the surface of the rotating body as the detection area, and set as the detection area; in the continuous image creation configuration, images on the plurality of temporary straight lines are extracted and combined in chronological order to create a plurality of continuous images; in the feature amount calculation step, the Hog feature amount is calculated for each of the plurality of continuous images; and in the rotation direction determination step, the rotation direction of the rotating body is determined based on the Hog feature amount of each of the plurality of continuous images, and the rotation direction of the rotating body is determined by majority vote based on the respective determination results (hereinafter referred to as the sixth method for determining the rotation direction of a rotating body according to the present invention).

[0020] According to the sixth method for determining the rotation direction of a rotating body of the present invention, the rotation direction of the rotating body is determined by majority vote based on the determination results for three consecutive images, thereby enabling more accurate detection and determination of the rotation direction of the rotating drum of a mixer truck that transports ready-mix concrete. [Effects of the Invention]

[0021] According to the present invention, by determining the rotation direction of a rotating body from the moving image information of the photographed rotating body, it is possible to quickly and accurately detect and determine the rotation direction of the rotating body shown in the image while suppressing the load on the computer. And it can be realized at low cost.

[0022] In addition, it is possible to quickly and accurately detect and determine the rotation direction of the rotary drum of a mixer truck that transports fresh concrete. It is possible to prevent the loading operation from being performed in the reverse rotation state.

Brief Description of Drawings

[0023] [Figure 1] It is a block diagram schematically showing the configuration of a rotating body rotation direction determination system in an embodiment of the present invention. [Figure 2] It is a block diagram schematically showing the electrical configuration of the rotating body rotation direction determination system of FIG. 1. [Figure 3] It is a conceptual diagram for explaining the setting of a detection area in the rotation direction determination process. [Figure 4] It is a diagram showing an example of a continuous image created in the rotation direction determination process. [Figure 5] It is a diagram (Part 1) for explaining a method of determining the rotation direction from a continuous image in the rotation direction determination process, and is a conceptual diagram for dividing a continuous image into a plurality of blocks. [Figure 6] It is a diagram (Part 2) for explaining a method of determining the rotation direction from a continuous image in the rotation direction determination process, and is a conceptual diagram for calculating a gradient direction histogram from a plurality of blocks. [Figure 7A] It is a diagram (Part 3) for explaining a method of determining the rotation direction from a continuous image in the rotation direction determination process, and is a diagram when detecting the rotation direction (forward rotation) with a histogram. ​​​​This is a diagram (part 5) illustrating how to determine the direction of rotation from a series of images in the rotation direction determination process, specifically showing the case where the direction of rotation (stop) is detected using a histogram. [Figure 8] Figure 1 is a flowchart illustrating the operation of the rotation direction determination system for a rotating body, from vehicle entry confirmation to the start of loading operations. [Figure 9] Figure 1 is a flowchart illustrating the operation of the rotation direction determination process in the rotation direction determination system for a rotating body. [Modes for carrying out the invention]

[0024] Hereinafter, embodiments of the rotation direction determination system and determination method for a rotating body according to the present invention will be described with reference to Figures 1 to 9.

[0025] Figure 1 schematically shows the configuration of the rotation direction determination system 100 for a rotating body in one embodiment of the present invention, and Figure 2 schematically shows the electrical configuration of the rotation direction determination system 100 for a rotating body. In this embodiment, the rotating body R is the drum of a mixer truck M that transports ready-mixed concrete.

[0026] In this embodiment, the rotation direction determination system 100 for a rotating body is configured using an existing surveillance camera system. As shown in Figure 1, it consists of a shooting means 10, a camera power supply unit 1, a distributor 2, a monitoring monitor 3, a converter 4, and a computer 5. The computer 5 contains the rotation direction determination system for the rotating body and a weighing operation management system. In this case, the distributor 2 and the converter 4 are newly added to an existing surveillance camera and weighing operation system.

[0027] In this rotating body rotation direction determination system 100, for example, the imaging means 10 is installed on the field side, and the computer 5 is installed in the control room.

[0028] As shown in Figure 2, the electrical configuration of the rotating body rotation direction determination system 100 consists of a computer comprising a display unit 101, an output means (transmitter) 102, a receiver 103, an input unit 104, a data bus 105, a storage unit 106, and a processing unit 107.

[0029] As shown in Figures 1 and 2, the rotation direction determination system 100 of this embodiment includes: an imaging means 10 for photographing the rotating body (drum) R to be determined; an image acquisition means 20 for acquiring a plurality of images at predetermined frame intervals from video information captured by the imaging means 10; a detection region setting means 30 for setting portions having characteristics on the surface of the rotating body R as detection regions to be used for detection in the plurality of images acquired by the image acquisition means 20; a continuous image creation means 40 for extracting the detection regions set by the detection region setting means 30 from the plurality of images and combining them in chronological order to create a continuous image; a feature amount calculation means 50 for calculating a Hog feature amount, which is a feature amount that is a histogram of the gradient direction of brightness, for the continuous image created by the continuous image creation means 40; and a rotation direction determination means 60 for determining the rotation direction of the rotating body (drum) R based on the Hog feature amount calculated by the feature amount calculation means 50.

[0030] The imaging means 10 is used to photograph the drum of the mixer truck M that transports ready-mix concrete. This imaging means 10 can be a surveillance camera. The imaging means 10 is also configured to transfer the digital image signal obtained from the image to the image acquisition means 20 via wired or wireless communication means.

[0031] The image acquisition means 20 acquires multiple images at predetermined frame intervals from the video information captured by the shooting means 10. In this embodiment, the image acquisition means 20 uses, for example, a video with 5 frames per second. 90 frames are used for one determination. The images acquired by the image acquisition means 20 are stored in the image data storage unit 106b.

[0032] The detection region setting means 30 sets a target region in a plurality of images acquired by the image acquisition means 20, specifying the portion of the rotating body (drum) R that has distinctive features as the detection region to be used for detection. In this embodiment, the detection region setting means 30 cuts out the same portion from each image with a single straight line L and sets it as the detection region. Here, the portion to be cut out is on the rotating body (drum) R, which is the object whose rotation direction we want to detect, as shown by the straight line L in Figure 3. Furthermore, the straight line L is set to be parallel to the rotation direction. In addition, the position of the straight line L is set to include the welded portion of the drum in order to facilitate the detection of features.

[0033] The continuous image creation means 40 extracts detection regions (straight lines L) set by the detection region setting means 30 from multiple images and combines them in chronological order to create a continuous image. In this embodiment, the continuous image created by the continuous image creation means 40 is shown in Figure 4. As shown in Figure 4, the cropped images were combined from top to bottom in chronological order to create a continuous image. In this case, the continuous image (grayscale processed) shows the characteristics shown due to dirt, scratches, etc. adhering to the rotating body (drum) R. Specifically, in Figure 4(A), forward rotation: an upward sloping line is detected. In Figure 4(B), stationary: it is detected as a vertical line. In Figure 4(C), reverse rotation: an upward sloping line is detected. The continuous image created by the continuous image creation means 40 is stored in the continuous image storage unit 106c.

[0034] The feature calculation means 50 calculates Hog features, which are features obtained by histogramming the direction of the brightness gradient, for the continuous images created by the continuous image creation means 40. In this embodiment, feature portions (slope, vertical lines) are extracted from the created continuous images. Hog features, which are features obtained by histogramming the direction of the brightness gradient, are calculated. The Hog features calculated by the feature calculation means 50 are stored in the feature calculation result storage unit 106d.

[0035] When calculating Hog features, first, as shown in Figure 5, the grayscale sequential image is divided into multiple blocks. Then, each block is divided into cells. Furthermore, it is divided into blocks (2x2 cells) and cells (3x3 pixels).

[0036] Next, let I(x,y) be the pixel of interest within the cell, and calculate the derivative Ix in the x-direction and the derivative Iy in the y-direction using the following equations (1) and (2).

number

number

[0037] The gradient strength |I|, which represents the strength of the gradient slope, and the gradient direction θ, which represents the angle of the gradient, are calculated from Ix and Iy using the following equations (3) and (4).

number

number

[0038] The feature calculation means 50 then performs the following steps: 1) quantizes the gradient direction in 19 directions (10 degrees each from 0 to 180 degrees). 2) calculates a gradient direction histogram for each cell, weighting it by intensity. 3) normalizes each block and calculates a gradient direction histogram for each block. 4) combines all the histograms to obtain the Hog features of the image. This is shown in Figure 6.

[0039] The rotation direction determination means 60 determines the rotation direction of the rotating body R based on the Hog features calculated by the feature calculation means 50. In this embodiment, when the rotation direction determination means 60 detects the rotation direction from the histogram, as shown in Figures 7A to 7C, it uses the difference between A "total number of change values ​​from 10 to 80 degrees" and B "total number of change values ​​from 100 to 170 degrees" from the histogram results as the determination result (see Figures 7A to 7C).

[0040] The left images in FIGS. 7A to 7C are continuous images created under each condition of the rotating body R. The right graph below is a histogram showing the relationship between the direction of the brightness transition created from the continuous images and the amount of change in each direction. FIG. 7A shows the case of detecting the rotation direction (clockwise rotation) using the histogram. That is, when A < B, it is considered clockwise rotation. FIG. 7B shows the case of detecting the rotation direction (counterclockwise rotation) using the histogram. That is, when A > B, it is considered counterclockwise rotation. FIG. 7C detects the rotation direction (stop) using the histogram. That is, when A ≈ B, it is considered stopped. The determination result by the rotation direction determination means 60 is stored in the rotation direction determination result storage unit 106e.

[0041] Regarding the electrical configuration of the rotating body rotation direction determination system 100, the main part of the computer 5 will be described with reference to FIG. 2.

[0042] The display unit 101 is composed of an LCD (Liquid Crystal Display) or an EL display (Electroluminescence Display), etc., and is configured to display message information, etc. composed of characters or images.

[0043] The output means (transmission unit) 102 constitutes the output means of the present embodiment, and is configured to transmit an output signal such as a determination result processed by the processing unit 107 to the outside via a wired or wireless communication network.

[0044] The receiving unit 103 is configured to receive information such as a digital image signal transmitted from the photographing means 10. The digital image signal (image data) acquired by this receiving unit 103 is stored in the storage unit 106. Further, the receiving unit 103 is configured to be able to receive a control signal from the outside via a wired or wireless communication network.

[0045] The input unit 104 is composed of a keyboard or a touch panel, etc., and is for inputting registration information, transmission / reception addresses, a processing operation start command, and other necessary information. Note that the input unit 104 may be connected and installed only when necessary.

[0046] The data bus 105 is connected between the display unit 101, the output means (transmitter) 102, the receiving unit 103, the input unit 104, the storage unit 106, and the processing unit 107, and is provided to enable data transfer between them according to the control program.

[0047] The storage unit 106 is mainly composed of, for example, a hard disk drive (HDD) and RAM and ROM memory. This storage unit 106 is equipped with a control program storage unit 106a, an image data storage unit 106b, a continuous image storage unit 106c, a Hog feature calculation result storage unit 106d, and a rotation direction determination result storage unit 106e. The control program storage unit 106a stores a program for controlling the operation of the rotation direction determination system 100 for a rotating body. The image data storage unit 106b temporarily stores multiple images acquired at predetermined frame intervals from image data captured by the shooting means 10. The continuous image storage unit 106c stores a continuous image created by extracting a set detection region from multiple images and combining them in chronological order. The feature calculation result storage unit 106d stores Hog features calculated by histogramming the luminance gradient direction for the continuous image. Furthermore, the rotation direction determination result storage unit 106e stores the result of determining the rotation direction of the rotating body based on the calculated Hog feature quantities.

[0048] The processing unit 107 is equipped with a CPU (Central Processing Unit) and controls the overall operation of the rotation direction determination system 100 of the rotating body according to a control program. This control program is configured to construct an image acquisition means 20 that acquires multiple images at predetermined frame intervals from video information captured by the shooting means 10, a detection area setting means 30 that sets the parts with distinctive features on the surface of the rotating body (drum) R in the acquired multiple images as detection areas to be used for detection, a continuous image creation means 40 that extracts the set detection areas and combines them in chronological order to create a continuous image, a feature amount calculation means 50 that calculates a Hog feature amount, which is a feature amount that is a histogram of the gradient direction of brightness, from the created continuous image, and a rotation direction determination means 60 that determines the rotation direction of the rotating body (drum) R based on the calculated Hog feature amount.

[0049] Hereinafter, a method for determining the rotation direction of a rotating body R using the rotation direction determination system 100 of this embodiment will be described with reference to Figures 8 and 9.

[0050] A method for determining the direction of rotation of a rotating body using the rotation direction determination system 100 comprises: an imaging step of imaging the rotating body (drum) R to be determined; an image acquisition step of acquiring multiple images at predetermined frame intervals from the captured video information; a detection region setting step of setting the parts on the surface of the rotating body (drum) R that have characteristics in the acquired multiple images as detection regions to be used for detection; a continuous image creation step of extracting the set detection regions from the multiple images and combining them in chronological order to create a continuous image; a feature calculation step of calculating a Hog feature quantity, which is a feature quantity that is a histogram of the gradient direction of brightness, from the created continuous image; and a rotation direction determination step of determining the direction of rotation of the rotating body (drum) R based on the calculated Hog feature quantity.

[0051] In this embodiment, the operation of the rotation direction determination system 100 from vehicle entry confirmation to the start of loading work will be described with reference to Figure 8.

[0052] As shown in Figure 8, first, on the field side, the entry of a mixer truck transporting ready-mix concrete is confirmed (step S1). If a mixer truck is found to be entering, the loading interlock is turned ON (step S2). Next, image processing is performed to determine the rotation direction of the mixer truck's drum (step S3). Then, based on the results of the image processing, it is determined whether or not the drum R of the mixer truck is rotating in the forward direction (step S4). If it is determined that the drum R of the mixer truck is rotating in the forward direction, the process proceeds to step S6. On the other hand, if it is determined that the drum R of the mixer truck is not rotating in the forward direction (i.e., if it is stopped or rotating in the reverse direction), in step S5, it is determined whether or not the loading release button is ON. If it is determined that the loading release button is not ON, the process returns to step S3 and the above processing is repeated until forward rotation is detected. On the other hand, if it is determined in step S5 that the loading release button is ON, the process proceeds to step S6. Here, the loading interlock is forcibly released by pressing the "loading prohibition release button" and loading begins. Then, in step S6, the loading interlock is turned OFF. Then, the loading of ready-mix concrete begins (Step S7).

[0053] Next, the operation of the rotation direction determination process in step S3 in Figure 8 will be explained with reference to Figure 9.

[0054] As shown in FIG. 9, when determining the rotation direction of the drum of a mixer truck using the rotation direction determination system 100 of a rotating body, first, a plurality of images (for example, 90 frames) are acquired at a predetermined frame interval (for example, 1 second) from the video information captured by the imaging means 10 (step S31). Next, it is determined whether the number of acquired images has reached the predetermined number of frames (step S32). Here, if it is determined that the number of frames has not reached the previously defined 90 frames, the process returns to step S31, and the above-described process is repeated. On the other hand, if it is determined that the number of frames has reached the previously defined 90 frames, in the plurality of acquired images, a detection region used for detecting a portion having a feature on the surface of the rotating body (drum) R is set (step S33). For example, the same portion is cut out with a straight line from each image (see FIG. 3). Next, a continuous image is created from the cut-out images (step S34). Here, the cut-out images are combined from top to bottom in chronological order to create a continuous image (see FIG. 4).

[0055] Next, for the created continuous image, a Hog feature amount, which is a feature amount obtained by histogramming the gradient direction of the luminance in a local region, is calculated (step S35). Here, feature portions (inclinations, vertical lines) within the created continuous image are extracted (see FIGS. 5 and 6).

[0056] Next, based on the calculated Hog feature amount, the rotation direction of the rotating body (drum) R is determined (step S36). Here, the rotation direction is detected from the histogram. If A < B, it is a forward rotation. If A > B, it is a reverse rotation. If A ≈ B, it is a stop (see FIGS. 7A to 7C).

[0057] In this way, by using the video information captured by the imaging means 10 such as a surveillance camera, the rotation direction of the drum of a mixer truck for transporting fresh concrete can be detected and determined quickly and accurately.

[0058] Hereinafter, the effects of the present embodiment will be described.

[0059] As described in detail above, according to this embodiment, the rotation direction determination system 100 for a rotating body comprises a shooting means 10, an image acquisition means 20, a detection area setting means 30, a continuous image creation means 40, a feature quantity calculation means 50, and a rotation direction determination means 60.

[0060] This system works by using the following steps to quickly and accurately detect and determine the rotation direction of the mixer truck drum as seen in the image: the shooting means 10 captures the rotating body (drum) R to be judged; the image acquisition means 20 acquires multiple images from the captured video information at predetermined frame intervals; the detection area setting means 30 sets the parts with distinctive features on the surface of the rotating body (drum) R in the acquired multiple images as detection areas to be used for detection; the detection areas set by the detection area setting means 30 are extracted from the multiple images; the continuous image creation means 40 combines them in chronological order to create a continuous image; the feature amount calculation means 50 calculates Hog features, which are features that are histograms of the brightness gradient direction, from the created continuous image; and the rotation direction determination means 60 determines the rotation direction of the rotating body (drum) R based on the calculated Hog features. This suppresses the load on the computer while quickly and accurately detecting and determining the rotation direction of the mixer truck drum as seen in the image. Therefore, it is possible to prevent loading work from being performed in the reverse rotation state. Furthermore, since it is configured using an existing surveillance camera system, it can be implemented at low cost.

[0061] In the embodiments described above, the detection region setting means 30 was shown to cut out the same portion from each image with a single straight line and set it as a detection region. However, the present invention is not limited to this. For example, the system may be configured to cut out the same portion from each image with multiple straight lines L and set each of them as a detection region.

[0062] When the detection area setting means 30 sets only one detection area (detection location), the drum rotation may not be detected due to differences in mixer truck models or stopping positions. Therefore, the rotation direction determination system 100 allows multiple detection lines (straight lines L) (for example, up to 3 lines) to be set as detection areas, thereby expanding the detection range for drum rotation direction and improving detection accuracy.

[0063] In the rotation direction determination system 100 for a rotating body, the detection method when using three lines involves measuring the difference value for each detection line and obtaining the detection result.

[0064] The final detection result is determined by majority vote from the three detection results (see Table 1). [Table 1] As shown in Table 1, the result of detection line 1 is "forward rotation", the result of detection line 2 is "forward rotation", and the result of detection line 3 is "stopped", therefore the final result is "forward rotation".

[0065] Furthermore, if all three detection results are different, the ratio is calculated from the absolute difference between the forward rotation and the reverse rotation. The ratio is calculated using the following formula. Ratio (%) = Absolute value of difference (smaller value) / Absolute value of difference (larger value) × 100

[0066] Here, if the ratio is less than 50%, no difference is observed in the absolute value of the difference between forward and reverse rotation, so it is determined to be "stopped". On the other hand, if the ratio is 50% or more, the result of the determination of the larger of the absolute values ​​of the difference is adopted. An example is shown in Table 2.

[0067] As shown in Table 2, the difference value (smaller value) is the absolute value of the difference in the reverse rotation, and the difference value (larger value) is the difference in the forward rotation. [Table 2] Here, the ratio is calculated: 2500 / 5000 × 100 = 50 Since the absolute value of the difference for forward rotation is 50% greater than the absolute value of the difference for reverse rotation, the final determination result is "forward rotation".

[0068] Furthermore, in the embodiment described above, the rotation direction determination means 60 may be configured to determine the effectiveness of the Hog features calculated by the feature calculation means 50 based on a preset threshold, and to determine the rotation direction of the rotating body R based on the Hog features that are determined to be effective.

[0069] As a result, in the rotation direction determination process, the effectiveness of the Hog features calculated by the feature calculation means 50 is determined based on a preset threshold, and the rotation direction of the rotating body R is determined based on the Hog features that are deemed effective. This allows for accurate determination by accommodating situations where the appearance of the rotating drum part differs depending on the surveillance camera footage, and where it is easier or harder to obtain difference values.

[0070] Furthermore, although the above-described embodiment explains an example in which the rotation direction of a mixer truck drum is determined as the rotating body R, the present invention is not limited to this. It can also be applied to determining the rotation direction of other rotating bodies.

[0071] Furthermore, although the above-described embodiments have included examples using existing surveillance cameras and weighing operation systems, the present invention is not limited thereto. A system independent of existing surveillance cameras and weighing operation management systems may also be configured.

[0072] The embodiments described above are all illustrative and not limiting, and the present invention can be implemented in various other variations and modifications. Accordingly, the scope of the present invention is defined solely by the claims and their equivalents. [Explanation of Symbols]

[0073] 1. Camera power supply unit 2 distributor 3. Surveillance monitor 4 Converters 5. Computer (PC) 10. Method of shooting 20 Image acquisition method 30 Detection area setting means 40. Continuous image creation method 50 Feature Calculation Method 60 Rotation direction determination means 100 Rotation Direction Determination System for Rotating Bodies 101 Display section 102 Output means (transmitter) 103 Receiving Unit 104 Input section 105 Data Bus 106 Storage section 106a Control program storage unit 106b Image data storage unit 106c Continuous image storage unit 106d Hog Feature Calculation Result Storage Unit 106e Rotation direction determination result storage unit 107 Processing Unit L Line (Detection Region) M Mixer truck R Rotating body (drum)

Claims

1. A means for photographing the rotating body to be judged, Image acquisition means that acquires multiple images at predetermined frame intervals from video information captured by the aforementioned shooting means, A detection region setting means sets a portion having a characteristic feature on the surface of the rotating body as a detection region to be used for detection in a plurality of images acquired by the image acquisition means, A sequential image creation means extracts detection regions set by the detection region setting means from the plurality of images and combines them in chronological order to create a sequential image, A feature calculation means calculates a Hog feature, which is a feature obtained by histogram-forming the luminance gradient direction, from the continuous image created by the continuous image creation means, A rotation direction determination system for a rotating body, characterized by comprising: a rotation direction determination means for determining the rotation direction of the rotating body based on the Hog feature calculated by the feature calculation means.

2. The aforementioned rotating body is the rotating drum of a mixer truck that transports ready-mixed concrete. The detection region setting means is configured to set a detection region by drawing at least one temporary straight line in a direction parallel to the direction of rotation at a location that includes a characteristic portion on the surface of the rotating body. The continuous image creation means is configured to extract images from at least one provisional straight line and combine them in chronological order to create at least one continuous image. The rotation direction determination system for a rotating body according to claim 1, characterized in that the feature calculation means is configured to calculate the Hog feature for at least one sequence of images.

3. The detection area setting means is configured to set a detection area by drawing a plurality of temporary straight lines in a direction parallel to the direction of rotation at a location that includes a characteristic portion on the surface of the rotating body. The continuous image creation means is configured to extract images from each of the multiple temporary straight lines and combine them in chronological order to create multiple continuous images. The feature calculation means is configured to calculate the Hog feature for each of the plurality of consecutive images. The rotation direction determination system for a rotating body according to claim 2, characterized in that the rotation direction determination means is configured to determine the rotation direction of the rotating body based on the Hog feature quantities of each of the plurality of consecutive images, and to determine the rotation direction of the rotating body by majority vote based on the respective determination results.

4. A shooting process to photograph the rotating body to be judged, Image acquisition process: Obtaining multiple images from captured video information at predetermined frame intervals, A detection region setting step involves setting a portion of the surface of the rotating body having distinctive features in multiple acquired images as a detection region to be used for detection, A sequential image creation step involves extracting a set detection region from the aforementioned multiple images and combining them in chronological order to create a sequential image. The process involves calculating Hog features, which are features obtained by histogramming the luminance gradient direction from the created sequence of images, and A method for determining the rotation direction of a rotating body, characterized by comprising a rotation direction determination step of determining the rotation direction of the rotating body based on the calculated Hog feature quantity.

5. The aforementioned rotating body is the rotating drum of a mixer truck that transports ready-mixed concrete. In the detection region setting step, at least one temporary straight line is drawn in a direction parallel to the rotation direction at a location that includes a characteristic portion on the surface of the rotating body, and this is set as the detection region. In the aforementioned sequential image creation step, images are extracted from at least one provisional straight line and combined in chronological order to create at least one sequential image. The method for determining the rotation direction of a rotating body according to claim 4, characterized in that the feature calculation step is performed to calculate the Hog feature for at least one sequence of images.

6. In the detection area setting step, multiple temporary straight lines are drawn in a direction parallel to the rotation direction at positions that include characteristic parts on the surface of the rotating body, and these are set as the detection area. In the aforementioned sequential image creation configuration, images are extracted from each of the multiple temporary straight lines and combined in chronological order to create multiple sequential images. In the feature calculation step, the Hog feature is calculated for each of the multiple consecutive images. The method for determining the rotation direction of a rotating body according to claim 5, characterized in that the rotation direction determination step is performed to determine the rotation direction of the rotating body based on the Hog feature quantities of each of the plurality of consecutive images, and then to determine the rotation direction of the rotating body by majority vote based on the respective determination results.

Images